A brief history of balancer chromosomes

The idea that heterozygosity for an inversion could suppress exchange was first proposed by Sturtevant [4] as a simple way to explain the observation that some chromosomes show a reduction of crossing over in a single region: “…individuals bearing one normal chromosome and one chromosome with an inverted section would probably show no crossing over in the region in question… .” As proof, he demonstrated that a crossover suppressor known as C_III [5] on Chromosome 3 was in fact an inverted segment later renamed In(3R)C [4,6]. This inversion, which involves the distal one-third of chromosome arm 3R, is present on most third chromosome balancers used today (Fig 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Commonly used balancer chromosomes in Drosophila showing approximate sizes of inverted segments.

X chromosome balancers invert an uncertain amount of pericentric heterochromatin (gray). Second and third chromosome balancer inversions do not bisect centric heterochromatin, and thus heterochromatin is not represented here. The large duplication present on SM5 is indicated (orange box).

https://doi.org/10.1371/journal.pgen.1008421.g002

Hermann Muller was likely the first to use the term “balanced” to describe enforced heterozygosity in a stock with a recessive lethal mutation on one chromosome and a crossover suppressor plus a recessive lethal on a homolog [7,8]. The first balanced stock involved In(3R)C, which carried a recessive lethal, in stock with a homolog carrying a recessive lethal allele of Serrate (Ser) with a dominant visible wing phenotype called Beaded (Ser^Bd-1) [7]. Had there been free recombination between In(3R)C and the homolog, Ser⁺ from In(3R)C would have replaced Ser^Bd-1, and the newly generated mutation-free chromosome would have quickly outcompeted both progenitor chromosomes to eliminate them from the stock population. By the 1920s, Muller had also characterized a large inversion of the middle of the X chromosome that carried several recessive visible markers, a dominant visible, and an unknown recessive lethal [9]. This chromosome, ClB, was an essential tool in the work that led to his Nobel prize in 1946. In(1)dl-49, a second X chromosome inversion discovered by Muller, inverts the middle third of the X, and although both chromosomes allowed infrequent proximal crossover events, In(1)dl-49 did not allow frequent double crossovers inside of the inversion as ClB did. Additional X chromosome inversions were generated by Muller using X-irradiation, including several whole-arm inversions that placed centric heterochromatin near the distal scute gene [10].

Because it was obvious that single inversions did not fully suppress exchange over the entire X, efforts were undertaken to build a more effective balancer. By the 1950s, a balancer combining In(1)dl-49 and a scute inversion existed and was known as First Multiple 1 (FM1) [11]. With two inversions, this chromosome functioned well as a balancer [12]. Further X-irradiation resulted in additional inversions, which prevented the proximal exchanges that could occur on FM1 [13,14]. One of the new balancers was FM7, the most commonly used X chromosome balancer today [15].

On Chromosome 2, Ward [16] was the first to describe two naturally occurring paracentric inversions that are the progenitors of nearly all second chromosome balancers used today. She was studying the Curly mutation isolated from a wild population and noted that when it was heterozygous, reductions in exchange occurred on both chromosome arms. Similar to the early X chromosome balancers, In(2L)Cy + In(2R)Cy was not an effective balancer for all second chromosome regions and was subjected to X-irradiation to introduce new inversions. A whole-chromosome inversion resulted in Second Multiple 1 (SM1) [11], with subsequent irradiation of SM1 resulting in SM5, which carried additional inversions and a large duplication [17]. Separate irradiation of In(2L)Cy + In(2R)Cy created a new balancer known as Curly of Oster (CyO) [18]. Single exchange events between SM1 and CyO later produced SM6 [19].

As noted above, third chromosome balancers also began as simple inversions that were combined and irradiated to produce more complex and effective balancers. X-irradiation of a third chromosome carrying two inversions, In(3LR)sep and In(3R)C, generated three new inversions to create Third Multiple 3 (TM3) [20]. A different chromosome carrying three existing inversions, In(3L)P, In(3LR)P88, and In(3R)C was irradiated to create TM6; subsequent exchange of the left arm of TM6 with another existing inversion, In(3LR)HR33 (which also carried the three-breakpoint inversion In(3R)Hu), yielded TM6B [21–23].

The creation of balancers required a sophisticated understanding of chromosome manipulation and stands as a testament to the genius of midcentury Drosophila geneticists. In contrast, balancing mutations on the small fourth chromosome is simple. It does not undergo meiotic recombination and therefore does not need a multiply inverted chromosome to suppress exchange. Any fourth chromosome carrying a recessive lethal or sterile mutation can effectively act as a fourth chromosome balancer, although usually a recessive lethal mutation with a dominant visible phenotype, such as eyeless^D, is used.

Balancers in other species

Any chromosome carrying at least one inversion and a closely linked recessive lethal or sterile mutation can function as a balancer for specific chromosomal regions. Increasing the number of inversions and rearrangements allows a chromosome to function as a balancer for more regions, and dominant visible markers assist in following the balancer in crosses, but they are not absolutely necessary. Because crossovers are suppressed in the vicinity of any heterozygous aberration breakpoint, translocations, transpositions, and duplications can also be used as balancers—though these uses are rare in Drosophila melanogaster.

Inversions have been identified and studied in several Drosophila species in addition to D. melanogaster. For example, several inversions have been described on the third chromosome of D. pseudoobscura, including some that overlap [24]. Inversions have also been well described for many other species, and stocks with either inversions or translocations for at least D. pseudoobscura, D. simulans, and D. virilis are available at the National Drosophila Species Stock Center (http://blogs.cornell.edu/drosophila/).

Outside Drosophila, chromosomes acting as balancers can be found in a handful of organisms. For example, in Caenorhabditis elegans, chromosomes carrying translocations, duplications, or inversions function as balancers [25–30]. Together, these chromosomes cover most, if not all, of the C. elegans genome, but unlike D. melanogaster, the majority of the C. elegans balancers suppress recombination over only small chromosomal regions near their breakpoints. Inversions that suppress exchange, some with recessive lethal mutations and some with dominant visible markers, also exist in Mus musculus, but they are available for only a small portion of the genome [31,32].

The hidden secrets of balancer chromosomes

Despite the pervasiveness of balancers in D. melanogaster research, the precise positions of the inversion breakpoints on the most commonly used balancers were only recently elucidated by whole-genome sequencing [33]. Forty-four of the 48 breakpoints on the most commonly used X, second, and third chromosome balancers were mapped using a combination of short-read Illumina sequencing and mate-pair sequencing. Subsequently, Ghavi-Helm and colleagues [34] used chromatin confirmation capture to estimate the positions of the four remaining breakpoints.

Although other balancers do exist in D. melanogaster, the balancers that have been sequenced are found in over 95% of the more than 40,000 stocks carrying at least one balancer at the Bloomington Drosophila Stock Center (https://bdsc.indiana.edu/stocks/stockdata.html). Furthermore, many of the inversions present on the balancers that were sequenced are also present on balancers that have not been sequenced.

Identifying the genomic position of each breakpoint yielded some surprising findings. Thirty-one breakpoints directly bisect a protein-coding gene, including the 65D breakpoint on TM3 that bisects all transcripts of the highly conserved tumor suppressor gene p53 [35]. Other genes have altered expression not because they are directly disrupted, but because of their proximity to a breakpoint. For example, the light gene, which encodes a cellular trafficking protein, is likely misexpressed on SM5 due to abnormal juxtaposition of euchromatic and heterochromatic regions at the 40F inversion breakpoint [36]. Whole-genome sequencing also determined that 117 protein-coding genes are present in a large duplication carried by SM5 (Fig 2).

Sequencing has shown that some balancers have been incorrectly labeled in stocks, and it is not difficult to see how that might happen. Most second chromosome balancers, for example, are marked with only one easy-to-identify dominant visible marker, Curly, leaving no simple way to determine if a stock carries CyO, SM1, SM5, or SM6. Indeed, in one study, out of 22 second chromosome balancers sequenced, four were mislabeled [36]. This should be a concern to researchers studying a mutant allele near a breakpoint or within a region poorly balanced by a particular balancer, such as the 42A to 58A segment of SM1. Fortunately, recessive markers can help distinguish balancers, and PCR primers are now available for 40 of the 44 breakpoints sequenced [35–37].

Another source of gene disruption on balancer chromosomes are SNPs and indels. Because each balancer was created one time and then distributed to the Drosophila community, any mutations on the original chromosome would be spread to all stocks containing that balancer. Because X-irradiation and ethyl methanesulfonate (EMS) mutagenesis were used to induce rearrangements and add visible markers, the number of mutations is probably higher on balancers. Indeed, sequencing a panel of balancers with common origins revealed many shared deleterious alleles, such as nonsense and splice-site variants. A panel of second chromosome balancers, for example, revealed 35 nonsense and 62 likely deleterious splice-site mutations shared among the all the stocks sequenced, as well as 8,898 missense variants whose impacts are unclear but may affect protein function [36]. Because balancers cannot easily replace deleterious alleles by crossing over, they are likely to accumulate unique mutations over time. One sequenced SM5 balancer, for example, was found to have one nonsense and 24 missense variants that were not found on the other four SM5 balancers sequenced [36].

In addition to the accumulation of SNPs and indels, sequencing revealed that balancers diverge in sequence as the result of rare double crossovers within inverted segments. For example, multiple X chromosome balancers had tracts of unique sequence within the 8.5-Mb In(1)dl-49 inversion, which lies in the middle one-third of the X chromosome. These novel tracts were introduced from their structurally normal-sequence homologs. All of these double crossovers replaced a female-sterile allele of the singed gene with a normal allele, which may have provided the new balancers with a competitive advantage in stock populations. Although crossovers between effective balancers and their homologs are rare, gene conversions appear to occur at rates similar to or higher than normal [1], providing a mechanism by which shorter tracts of new sequence can be introduced.

What balancers can teach us about crossover suppression by breakpoints

The distance over which inversion breakpoints suppress exchange is unknown and has been challenging to study using traditional marker-based approaches. Whole-genome sequencing of balancers has helped us chip away at this question. Because a large region at the distal end of left arm of TM3 can be exchanged with normal-sequence homologs by single crossovers, historical recombination events have been preserved in this interval. Sequencing several TM3 stocks revealed exchange events as close as 2 Mb from the distalmost inversion breakpoint, providing the first direct observation of the closest distance to a breakpoint a crossover can form in the face of crossover suppression [35].

In a subsequent study [1], crossovers were seen approximately 1 Mb from balancer breakpoints, but, interestingly, gene conversions were observed evenly distributed along the chromosomes and in close proximity to the breakpoints. This tells us that breakpoints have no effect on the placement of double-strand breaks or their repair into gene conversions. Considering data showing that gene conversions do not respond to interference or to the inhibition of meiotic recombination seen near centric heterochromatin, known as the “centromere effect” [33], it appears that inversion breakpoints may suppress exchange through a mechanism similar to interference or the centromere effect.

The future of balancers

The coming years will likely see precise changes made to balancers using new tools, such as CRISPR and transgene technologies, either to make structural modifications to enhance their function or to introduce alternative markers. Indeed, GFP-expressing transgenes have been added to balancers to speed up screening in mutagenesis experiments by, for example, allowing for the easy identification of balancer heterozygotes that might be selected for or against [38–40]. Multiply rearranged chromosomes will likely be created for other species of Drosophila, allowing the maintenance of deleterious alleles and experiments involving complex crosses. Balancer chromosomes have a rich history and have been key to the development of D. melanogaster as a prominent model organism, and it is clear they will be essential to the Drosophila community for many years to come.

Acknowledgments

We thank Angela Miller for assistance with figure preparation and editing.